stage2: detect redundant C/C++ source files

src/Cache.zig

@@ -26,6 +26,7 @@ pub fn obtain(cache: *const Cache) Manifest {
 /// This is 128 bits - Even with 2^54 cache entries, the probably of a collision would be under 10^-6
 pub const bin_digest_len = 16;
 pub const hex_digest_len = bin_digest_len * 2;
+pub const BinDigest = [bin_digest_len]u8;
 
 const manifest_file_size_max = 50 * 1024 * 1024;
 
@@ -41,7 +42,7 @@ pub const File = struct {
     path: ?[]const u8,
     max_file_size: ?usize,
     stat: fs.File.Stat,
-    bin_digest: [bin_digest_len]u8,
+    bin_digest: BinDigest,
     contents: ?[]const u8,
 
     pub fn deinit(self: *File, allocator: *Allocator) void {
@@ -139,16 +140,16 @@ pub const HashHelper = struct {
         return copy.final();
     }
 
-    pub fn peekBin(hh: HashHelper) [bin_digest_len]u8 {
+    pub fn peekBin(hh: HashHelper) BinDigest {
         var copy = hh;
-        var bin_digest: [bin_digest_len]u8 = undefined;
+        var bin_digest: BinDigest = undefined;
         copy.hasher.final(&bin_digest);
         return bin_digest;
     }
 
     /// Returns a hex encoded hash of the inputs, mutating the state of the hasher.
     pub fn final(hh: *HashHelper) [hex_digest_len]u8 {
-        var bin_digest: [bin_digest_len]u8 = undefined;
+        var bin_digest: BinDigest = undefined;
         hh.hasher.final(&bin_digest);
 
         var out_digest: [hex_digest_len]u8 = undefined;
@@ -241,7 +242,7 @@ pub const Manifest = struct {
         const ext = ".txt";
         var manifest_file_path: [self.hex_digest.len + ext.len]u8 = undefined;
 
-        var bin_digest: [bin_digest_len]u8 = undefined;
+        var bin_digest: BinDigest = undefined;
         self.hash.hasher.final(&bin_digest);
 
         _ = std.fmt.bufPrint(&self.hex_digest, "{x}", .{bin_digest}) catch unreachable;
@@ -347,7 +348,7 @@ pub const Manifest = struct {
                     cache_hash_file.stat.inode = 0;
                 }
 
-                var actual_digest: [bin_digest_len]u8 = undefined;
+                var actual_digest: BinDigest = undefined;
                 try hashFile(this_file, &actual_digest);
 
                 if (!mem.eql(u8, &cache_hash_file.bin_digest, &actual_digest)) {
@@ -381,7 +382,7 @@ pub const Manifest = struct {
         return true;
     }
 
-    pub fn unhit(self: *Manifest, bin_digest: [bin_digest_len]u8, input_file_count: usize) void {
+    pub fn unhit(self: *Manifest, bin_digest: BinDigest, input_file_count: usize) void {
         // Reset the hash.
         self.hash.hasher = hasher_init;
         self.hash.hasher.update(&bin_digest);
@@ -530,7 +531,7 @@ pub const Manifest = struct {
         // cache_release is called we still might be working on creating
         // the artifacts to cache.
 
-        var bin_digest: [bin_digest_len]u8 = undefined;
+        var bin_digest: BinDigest = undefined;
         self.hash.hasher.final(&bin_digest);
 
         var out_digest: [hex_digest_len]u8 = undefined;

src/Compilation.zig

@@ -33,6 +33,7 @@ gpa: *Allocator,
 arena_state: std.heap.ArenaAllocator.State,
 bin_file: *link.File,
 c_object_table: std.AutoArrayHashMapUnmanaged(*CObject, void) = .{},
+c_object_cache_digest_set: std.AutoHashMapUnmanaged(Cache.BinDigest, void) = .{},
 stage1_lock: ?Cache.Lock = null,
 stage1_cache_manifest: *Cache.Manifest = undefined,
 
@@ -1152,6 +1153,7 @@ pub fn destroy(self: *Compilation) void {
         entry.key.destroy(gpa);
     }
     self.c_object_table.deinit(gpa);
+    self.c_object_cache_digest_set.deinit(gpa);
 
     for (self.failed_c_objects.items()) |entry| {
         entry.value.destroy(gpa);
@@ -1730,6 +1732,17 @@ fn updateCObject(comp: *Compilation, c_object: *CObject, c_comp_progress_node: *
         }
     }
 
+    {
+        const gop = try comp.c_object_cache_digest_set.getOrPut(comp.gpa, man.hash.peekBin());
+        if (gop.found_existing) {
+            return comp.failCObj(
+                c_object,
+                "the same source file was already added to the same compilation with the same flags",
+                .{},
+            );
+        }
+    }
+
     var arena_allocator = std.heap.ArenaAllocator.init(comp.gpa);
     defer arena_allocator.deinit();
     const arena = &arena_allocator.allocator;